of Small Resistances, 293 



and in the case in which the resistances to be measured are 

 very nearly equal we may put P = Q, P=S, in the coefficient 

 of e, so that 



H + A{ 2B (W) +4 } • (2 > 



Thus if B is large compared with R and P, the term involving 

 e is appreciable, and the usual condition P : Q = R : S is not 

 sufficiently nearly true to give correct results. 



The equation shows that the sign of the correction changes 

 with that of E. 



Let F / and Q' be the values of P and Q when the battery 

 resistance is large. 



Then 



•"rWMWM- 



Q 



and 



e = 



E (!~9 



2Br+^+4 



<w> 



(S) 



With the view of testing the truth of this explanation, I 

 made a series of experiments, using a wire bridge and various 

 small resistances. 



The resistance of the bridge-wire in the experiments was 

 •08 ohm ; the wire was a metre long, and graduated to milli- 

 metres. 



Different observations for the value of P agree to a fraction 

 of a millimetre of the bridge-wire, while the differences in 

 the values of P' never amounted to more than 4 millimetres, 

 and were usually much less. The value of p was 150 ohms. 

 In some of the experiments Leclanche cells were used, in 

 others Daniells. The E.M.F. of the Leclanche cells was 

 found by the potentiometer method to be about 1*25 Daniell. 



In the results given in the table, the E.M.F. of a Daniell 

 cell is taken as 1 volt. The battery resistance is neglected 

 compared with the 150 ohms interposed. 



Phil Mag. S. 5. Vol. 11. No. 68. April 1881. 



